2020 Robotics Engineering Virtual MQPs

A two-day interactive opportunity for faculty judges, fellow students, and outside guests to listen to, learn from, and appreciate the work of our 2020 Major Qualifying Projects.

2020 MQP Projects

MQP 1 – Humanoid Animatronic Learning Simulator for Medical Interactive Training (H.A.L. S.M.I.T.)

The goal of this project was to design and develop an autonomous, modular, easily operable animatronic head to enhance realism in medical simulations. The technology is expensive for many medical programs, and manikins have limited capacity for complex tasks. This project addressed such gaps by producing a low cost head and neck manikin capable of performing complex simulations. A previous prototype was adapted by addressing areas of improvement for medical use. The target capabilities for the manikin included airway management, and level of consciousness and circulatory assessment. Ultimately, subsystems were tested and a final design was produced with the potential for additional functionality and wide applicability to the medical field.

MQP 2 – Active Telepresence Assistance for Supervisory Control: A User Study with a Multi-Camera Tele-Nursing Robot

Supervisory control of a humanoid robot in a manipulation task requires coordination of remote perception with robot action, which becomes more demanding with multiple moving cameras available for task supervision. We explore the use of autonomous camera control and selection to reduce operator workload and improve task performance in a supervisory control task. We design a novel approach to autonomous camera selection and control, and evaluate the approach in a user study which revealed that autonomous camera control does improve task performance and operator experience, but autonomous camera selection requires further investigation to benefit the operator’s confidence and maintain trust in the robot autonomy.

MQP 3 – Next Gen of Wearable Robotics

The purpose of this project was to create a prosthetic wrist that utilizes Hydro Muscles to actuate motion to complete simple tasks. This project consisted of the design of a wrist joint with two rotational degrees of freedom, an attachment to mimic the gripping motion with Hydro Muscle actuated fingers, a compact pump actuation system to convert rotational movement from a small motor to linear expansion of a Hydro Muscle, and the development of controls and feedback to control the actuation. Goals for this project were centered on ideas of controlled usage scenarios and real life implementation.

MQP 4 – Car-Snow Clearing Drone

The purpose of this MQP was to research, design, analyze, and test a robotic device to remove snow and ice off of cars. Snow on cars is a hazard for many drivers across the world. Drones are widely used in society, including agricultural drones for spraying pesticides on crops. Lightweight, yet strong, carbon fiber was used in addition to 3-D printed parts. The flying and spraying components were each tested in addition to proving the concept with multiple FEA simulations on individual components. Computer vision was used to identify how much snow remained on the car and what areas were already sprayed. The primary focus of the project was to come up with a device to aid society and potentially sell to consumers.

MQP 5 – The M.O.L.E. System: Autonomous, Trenchless Tunneling Robot

Digging tunnels is a necessary part of many construction and infrastructure projects, but most small-scale tunnel-boring methods presently require a trench which can disrupt aspects of daily life on the surface. There are trenchless methods, but they require that the tunnel be completely or almost completely straight. The goal of this project was to design a self-correcting machine capable of producing complex tunnels without the use of a trench. Early stages of design considered many different alternatives for steering, anchoring, and boring. A prototype was constructed as a proof of kinematic concept with full electronic and computational systems onboard. Using the prototype, a report was generated detailing the improvements that would need to be made for a high-budget and marketable iteration of the device because, while the prototype is not capable of travelling underground, it provides proof that the kinematic designs are sound.

MQP 6 – Telenursing RoboPuppet 2019-20

RoboPuppet is a scale-model arm of ReThink Robotics’ Baxter used for intuitive remote control. The arm contains joint angle sensors and motors which allow for gravity compensation and limited haptic feedback. The project includes a ROS package for controlling Baxter in real life and simulation, with a basic real-time GUI for calibration and debugging. This platform is ideal for helping nurses work remotely with patients in high-risk and contaminated environments.

MQP 7 – Demining Autonomous System

The use of PMN-1 anti-personnel landmines, along with poor documentation and scale of their deployment, has led to a humanitarian crisis. Current methods of landmine removal, including military-grade equipment, trained animals, and manual deminers, are dangerous, ineffective, and expensive. This project built on the work of previous MQPs to develop a more robust rover for autonomous landmine detection and a more accurate drone for landmine detonation. Additionally, the team developed an easy-to-use application that combines the two systems and provides access to non-technical users.

MQP 8 – Ultrasound Sensing for Prosthetic Control

Amputation in the arm results in decreased capability for individuals to manipulate their surroundings effectively. Advances in active prostheses have sought to augment these individuals with a prosthetic that returns some, but ideally all, functionality back to the individual. Specifically in transradial amputations, many of the muscles required to move the hand are still intact. Ultrasound can be used to produce a cross-sectional image of the forearm and the muscles present. Through the utilization of ultrasound images, assessments can be made on the muscles state and their effect on the downstream finger movements. Building off of prior work done in this field, this project aims to demonstrate use of ultrasound in prosthetic control. To accomplish this, we designed and analyzed a prosthetic design to implement this technology, built the prosthetic, and performed an assessment on the optimal probe orientation on the forearm.

MQP 9 – SCREAM 2.0: Superelastic Continuum Robot for Endoscopic Articulation and Manipulation

The objective of this project is to investigate robotic platforms to enhance a physician’s dexterity in minimally invasive endoscopic surgery. We present the design and construction of a continuum robotic manipulator consisting of two concentric notched tubes capable of providing pan and tilt degrees of freedom. Building the manipulator out of tubes enables the passage of a variety of surgical instruments through the tubes’ open lumen. The robot is equipped with a quick release mechanism which enables hot swapping of instruments during a procedure. Positional control of the robot is based on its inverse kinematics, which is approximated by means of a Newton-Raphson method combined with a geometric approach to generate an initial solution. Using an initial guess close to the desired solution ensures fast convergence. We report on experimental evidence that verifies the positional accuracy of the robot. Out of the many applications where we believe our robot would be helpful, one where it would help immensely is the endoscopic treatment of laryngeal tumors, which affect 1 in 40 people worldwide at any given time. These tumors are challenging to treat endoscopically due to the lack of articulation in commercially available surgical instruments. Other potential applications include ear surgery and kidney surgery.

MQP 10 – WPI Mascot Robot

Robotics is a signature program of WPI, however, there is still an opportunity to further integrate robotics into WPI’s image. While it is not the only important part of WPI’s identity, the robotics work its students and faculty produce are iconic and recognizable to people both in and out of STEM fields. The goal of this project is to design a prototype of a mascot robot for WPI that anyone can interact with to see and feel what WPI is all about. The aim of this project is not to replace Gompei, but to add to the Gompei experience with a little robot flair, as is WPI tradition.

The WPI Robot Mascot MQP is designing and building a robot goat to be a companion mascot to Gompei. Robotics is a signature program of WPI, and this MQP seeks to further represent and demonstrate the iconic robotics work done by both students and faculty. It would be impossible for a single MQP to develop a robot mascot that reflects the impressive robotics achievements at WPI. This preliminary MQP involved design and prototyping work for 5-DoF head and neck assembly. This includes work on the mechanical, electrical, and software systems of the robot, as well as the initial requirements and project planning done in collaboration with various stakeholders at WPI. Future MQP teams will further develop the robot, and it will serve as a continuously evolving platform to showcase the traditions and experiences at WPI.

MQP 11 – A Stride Towards the Elimination of Consumer Waste: Development of a Reusable Cup Machine

This project pursues an autonomous solution for reducing single use cup waste through a reusable cup sharing kiosk. A design concept suitable for this type of program must receive a dirty cup from a consumer and return a clean cup ready-for-use. Such a device must have washing, drying, storing, and dispensing capabilities. The design process consists of three main steps: brainstorming, analyzing, and comparing with a design matrix. The final product prioritizes compactness and simplicity of mechanical systems. Though significant progress was made in the implementation of the proposed design, there is still considerable work to be done to finish this prototype due to the interruption of the semester by the COVID-19 pandemic.

MQP 12 – DR-Swarm2

Robots are envisioned to work alongside humans. However, humans struggle to interpret the state and goals of a robot. The use of multiple robots further exacerbates this issue. To solve this problem, we propose Dr. Swarm 2, an augmented reality (AR) application built on the Magic Leap. The application conveys robot states and goals through virtual artifacts overlaid with the real world. The overlay provides concise information in a manner un-achievable with existing methods.

MQP 13 – Chemical Synthesis Robot MQP

The Chemical Synthesis MQP has developed a robotic platform that can automate parts of a chemical synthesis process. For purposes of this MQP, a chemical synthesis consists of reaction, separation, concentration, chromatography, and isolation. Our robotic platform consists of an interconnected network of client lab instruments that are driven by Arduino microcontrollers. In turn, these clients are controlled from a main server that is running LabVIEW. This server converts predefined chemical functions into a sequence of steps that are distributed to the corresponding clients to be executed by the lab instruments. During this MQP, we focused on developing an organic-aqueous separator and a reagent dosing system that incorporates a temperature feedback loop.

MQP 14 – NASA Lunabotics: LOADER

We have designed an autonomous mining robot designed to meet the specifications of the 2020 NASA robotic mining competition. The rover is designed to navigate in a moon simulated environment, dig down at least thirty centimeters and return a payload, as well as meet the constraints imposed by the unique operating environment. We addressed the engineering challenges and goals that were faced, ranging from communication and obstacle avoidance to heat management.

MQP 15 – INNDiE: An Integrated Neural Network Development Environment

Modern machine learning methods are capable of tackling problems that are traditionally difficult or impossible for computers to solve. These methods present a steep learning curve with a wall of information and concepts for novices and often requires expensive computing resources to implement. INNDiE solves these problems by helping the user configure, train, and test neural networks with limited programming and machine learning background knowledge via a graphical user interface. INNDiE also supports training neural networks in the cloud which reduces the cost of training a network.

MQP 16 – Robotic Water Monitoring

With devastating funding cuts propagating through environmental agencies countrywide, the monitoring of local water bodies has been one of the essential services to be curtailed. The health of every body of water is crucial to the diverse aquatic ecosystems as well as the human populations who reside near their shores. Poor water quality has a disastrous impact on a region much larger than the body's footprint, and changes are rarely reversible. Although sampling is critical, the process is time and resource intensive meaning most bodies of water are either unmonitored, or sampled by volunteer groups. The Robotic Water Monitoring boat is a tool to expedite sampling, enabling the state to better allocate scarce resources by significantly increasing the quantity and quality of water samples.

MQP 17 – Modular N-Link Smart Robotic Arm

Today, 2.6 million robots have been deployed worldwide for industrial applications, allowing us to produce, sort, and create products faster. Often, these robots require maintenance every 3,850 hours and have downtimes that yield production resulting in loss of efficiency and increase the cost. The goal of this MQP is to develop a modular robotic arm with n-links that can position its end effector in a 3D task space depending on the number of links attached. It is also a priority that the arm is fault-tolerant and can be removed and attached in real-time without requiring a reboot. Creating this arm will demonstrate how a modular robotic arm can operate in a dynamic task space by adding various links and decrease malfunction maintenance time with the improvement of fault tolerance.

MQP 18 – Providing an Assistive Torque at the Knee

This project aimed to develop an active assistive knee device to aid the user in their gait cycle and vertical propulsion. Currently, only several expensive, bulky solutions for people with knee ailments exist. Experimental and literary research was performed to determine the amount of assistance needed from the device, and the ideal structure. A novel hinge design was created and, once partnered with the electromyograph (EMG) and motor, will provide 10% of the required torque for the user. Improvements include refining the design to be smaller and lighter, and considering the addition of a second motor for increased torque.

MQP 19 – Lionfish Phase 3

Native to the Indo-Pacific region, Lionfish are an invasive species in the Carribean Sea that were first introduced by aquarium owners irresponsibly releasing their pets into the ocean. The lionfish population continues to threaten marine life in the region and WPI has aimed to address the problem. In its third year, this MQP proposes a fully autonomous robotic solution to kill and capture the invasive species. Building off of previous teams’ work, this year’s focus is on developing a new harvesting system, refined identification algorithm, and autonomous navigation algorithm. These systems will be integrated into an off-the-shelf ROV to create a completely self-contained Lionfish capturing robot.

MQP 20 – Modular Framework for Socially Assistive Robots 

Socially assistive robots (SARs) are designed to assist in the development of social skills. One promising application of SARs is in applied behavior analysis (ABA) therapy, which helps those diagnosed with autism spectrum disorder (ASD) develop social skills. To assist in ABA therapy, an SAR titled Penguin for Assistive Behavioral Intervention (PABI) was created. PABI uses modular social assistive robot framework (MSARF) to allow for further development and large-scale utilization of this innovative assistive technology. PABI is an essential ABA therapy tool because it performs rigid assessments and real-time data logging, which increases the quality of ABA therapy by reducing time spent manually logging data and enforcing strict adherence to therapy protocol. By implementing a MSARF robot that specializes in ABA therapy, the efficacy of PABI's framework can be tested.

MQP 21 – 3D Printer Robotic Bed Removal

The use of desktop 3D printers has increased significantly in recent years as quality has improved and prices have dropped. Even as 3D printing has become more accessible, users are still bottlenecked by needing to be present to remove parts from the printer bed and manually start the next print. Our project has removed this limitation by designing and building an autonomous system to remove parts from a printer and manage an online queue of jobs. These components form a reliable, user-friendly system that increases the utility and efficiency of a 3D printer.

MQP 22 – Scalable Sort Automation

Effective sortation and separation of useful components from bulk mixtures is a grand challenge in recycling and many other industries, but becomes increasingly challenging as the class count of the components being sorted increases. The age-old challenge of sorting LEGO, with a well-documented, highly diverse catalog of over 70,000 part variants, serves as an excellent test bed for developing sortation technologies. The aim of the project is to develop an effective and scalable automated LEGO sortation system, with a particular focus on developing a system architecture which scales economically as the range of items being sorted is increased. This ensures that key cost-driving parameters such as size, actuator count, or cycle time do not grow at an unacceptable rate relative to component class count, allowing for the practical implementation of systems capable of uniquely classifying and separating potentially tens of thousands of unique component classes. The proposed solution consists of a multi-tiered mechanical distribution system incorporating approaches such as grate sieving and conveyor serialization, with the goal of enabling continuous component flow. The mechanical system operates alongside a machine learning classification algorithm which incorporates approaches such as consensus protocols between competing algorithms to prioritize uncertainty detection and contamination prevention above absolute classification accuracy in the face of a challenging multi-class data set.

MQP 23 – Firefighting Remote Exploration Device II

The advancing technologies of today are increasing, and the need for “Smart” recovery for disasters is at the forefront. Firefighters operating in indoor firegrounds are put at risk by the constantly changing environment. The use of robotics in firefighting can assist firefighters by informing them about different aspects of the fireground, such as the structural layout and temperature distribution. Taking inspiration from a design devised by a previous WPI Major Qualifying Project, our team prototyped a heat, water, and impact-resistant robot capable of navigating around obstacles in the fireground and returning relevant real-time data.

MQP 24 – 3D Swarm Construction

Robotic construction can drastically improve the efficiency and safety of construction. However, current robotic construction is limited by the types of structures robots can build and the ability for multiple robots to work collaboratively to build structures. This project aims to create an autonomous collective construction system in which two types of robots cooperate: construction robots and smart scaffolding robots. The latter robot type integrates electronics into building materials to create intelligent structures and allows for dynamic reassembling of existing components. In addition, we design a multi-robot collaborative building algorithm that showcases construction both with real and simulated robots. This project is developed so as to establish both a physical and simulated framework for future work with swarm construction algorithms.

MQP 25 – Red Legs Robot

The Red Legs robot was created in 2011 to be a sensitive, bipedal walking platform. However, over the past few years, Red Legs has been neglected and has fallen into a state of disrepair. The goal of this project is to design a working computational, electrical, and mechanical system to enable the Red Legs robot to carry out basic movements and set up an intuitive framework for future work. The project will also aim to create comprehensive documentation along the way. To achieve this goal, the robot will be examined and studied to find out precisely its current state, and the manner in which it works. The project will also attempt to improve the designs of existing parts and software, and make necessary changes to allow for more complex movements.

MQP 26 – Robotic Waste Sorting

The recycling industry is struggling under tight profit margins, changing waste policies, and fast evolving waste market regulations. One notable issue is processing the highly contaminated single-stream recycling waste generated by the world’s growing population and consumption practices. To help provide a solution, this project contributes to the long-term goal of developing a waste-sorting robot to efficiently sort single-stream recyclables. Such technology will also change the role of recycling workers from handling the hazardous, and sometimes toxic waste streams to being robot and process supervisors. Robotic solutions would allow faster and more precise sorting of the waste streams, reducing the percentage discarded but essentially recyclable. Enabling the processing of currently discarded materials can also result in a significant decrease in landfill use. Our goal for this project was to develop a test bed suitable to simulate a real-world recycling plant environment containing a robot capable of identifying and removing cardboard from a stream of mixed recyclables.

The test bed we developed consists of three major components: the frame, the arm, and the control system. At the conclusion of our project, we succeeded in constructing an X-Z Cartesian platform together with its rail system and steel frame housing, a five-joint linkage arm that is mounted on the Cartesian platform, and a three-jaw gripper designed for effective cardboard picking. The control system was designed and successfully tested. A software stack was selected and the control abstraction functions were developed as a base for the future applications to build upon. Such outcomes provide the main architecture for establishing a robotic waste sorting experimental setup at WPI.

MQP 27 - Soft Psychophysiology

The goal of this project is to prospectively increase the accessibility of psychophysiological studies by using inexpensive, effective sensor technology from soft robotics. The growing field of soft robotics has developed novel sensor technology that is uniquely suited for unobtrusive measurement of movement data on the human body. In parallel, the field of psychophysiology is currently limited in available data-collection methods, which is a high barrier to entry for prospective studies. Some current methods quantify emotions like stress by using neurological and physical activity as proxies. This project aims to develop a relatively inexpensive device, utilizing soft robotics sensors, for the purpose of measuring stress on the human body.

MQP 28 – Beach Bots

Beaches across the globe are being polluted by human trash. The impact of beach pollution is an environmental hazard. If trash is continually removed from these beaches, it could result in a decreased threat to wildlife. While there are solutions to this problem, they mostly entail large machines that are not environmentally friendly and require a human operator. In its first year, this MQP aims to develop a robotic solution to clean these beaches effectively. As the project is in its first year, it will aim to develop building blocks for later MQP including the development of one small robot, a larger robot and some of the communication methods between them.

MQP 29 – Autonomous Navigation for Planetary Robots

Two robots can successfully navigate unknown terrain to find and recover targets whose global positions are known with the aid of a onetime method that returns the current position and orientation of the given robot. To achieve autonomous navigation, the robots must take sensory data from their surroundings, use the data to form an observable map, distinguish routes that can be safely traveled, and maintain knowledge of their current positions in respect to the map they have created.

MQP 30 – Sailbot 2019-2020

The goal of this project was to improve in major ways upon the autonomous robotic sailboat known as the “The Wide Awake.” Using the rules of the International Robotic Sailing Regatta as a guide, the team improved the boat’s mechanical, electrical, and software performance and reliability by improving upon the systems that were already in place from previous years as well as devising new solutions to existing problems. The focus of this year’s MQP was to implement reliable RC control and to pave the way for the implementation of autonomous operation in future years.

MQP 31 – Modular Self-Driving and Sensor Packages for R/C Cars

In this project, we created a modular self-driving kit and a sensor kit that can be integrated into any scale RC car. The self driving kit works by processing images taken from the car through a neural network to adjust the car’s steering and throttle. The sensor kit records the speed of drive shafts, rotations around pitch and roll axes, and temperature of the motor’s batteries and sensor boards as the car is driving. These kits will be used in WPI’s Advanced Engineering Design course where students design scale cars. The modular self driving kit and sensor kit will provide students with exposure to real-time data collection, and enhance their understanding in both mechatronic systems and autonomous driving developments.